It's really nice to run a dozen powders through a system while adjusting seating depth and hardness and observe muzzle pressure, pressure curve, peak pressure, peak pressure location, percent powder burn, muzzle velocity, and barrel time......and never leave my chair.
In the real world Quickload has been spot-on with velocity numbers (the only real way I have to check the predictions), so I assume the rest is pretty close. The program is absolutely invaluable for estimating safe starting loads (together with any other shreds of valid data that may be available, and a healthy dose of Handloader's Sense), and like Brad wrote I wouldn't rely on it for absolute safe maximums, but with a good chronograph I feel reasonably confident in pressure predictions if the MV numbers match predictions.
Cylinder gap and gas system pressure losses are so far the only things I've found that really trip up the Quickload predictions, as one would expect. Figuring things like gas port pressure in any gas-operated automatic help determine what powders and minimum loads will work before wasting any components. A little outside math applied using some of the heat and expansion data numbers from QL can give a clue of gas volume actually going through the port when you consider time of bullet in the barrel. You can do a lot with QL, but it's not a program for "dummies", it's a program for programmers, sort of like CNC programs and not at all like Windows.